fdurastante/psblas3
1
0
Fork 0
mirror of https://github.com/sfilippone/psblas3.git synced 2026-10-17 03:12:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

Update docs for allocate/deallocate _wrk

Browse Source
This commit is contained in:
sfilippone
2025-02-24 09:56:35 +01:00
parent 40f683effd
commit 65aac92c16
13 changed files with 9532 additions and 8819 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/html/userhtmlse13.html
+44 -44
View File
@@ -16,16 +16,16 @@ href="userhtmlse12.html#tailuserhtmlse12.html" >prev-tail</a>] [<a
href="userhtmlse10.html#tailuserhtmlse13.html">tail</a>] [<a
href="userhtml.html# " >up</a>] </p></div>
<h3 class="sectionHead"><span class="titlemark">13 </span> <a
id="x20-15500013"></a>CUDA Environment Routines</h3>
id="x20-15700013"></a>CUDA Environment Routines</h3>
<!--l. 91--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-156000"></a>psb_cuda_init &#8212; Initializes PSBLAS-CUDA environment</h4>
id="x20-158000"></a>psb_cuda_init &#8212; Initializes PSBLAS-CUDA environment</h4>
<a
id="Q1-20-194"></a>
id="Q1-20-196"></a>
<div class="center"
>
<!--l. 99--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-110">
<div class="minipage"><pre class="verbatim" id="verbatim-112">
call&#x00A0;psb_cuda_init(ctxt&#x00A0;[,&#x00A0;device])
</pre>
<!--l. 103--><p class="nopar" > </div></div>
@@ -64,17 +64,17 @@ class="pplb7t-x-x-120">Notes</span>
<ol class="enumerate1" >
<li
class="enumerate" id="x20-156002x1">
class="enumerate" id="x20-158002x1">
<!--l. 125--><p class="noindent" >A call to this routine must precede any other PSBLAS-CUDA call.</li></ol>
<!--l. 129--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-157000"></a>psb_cuda_exit &#8212; Exit from PSBLAS-CUDA environment</h4>
id="x20-159000"></a>psb_cuda_exit &#8212; Exit from PSBLAS-CUDA environment</h4>
<a
id="Q1-20-196"></a>
id="Q1-20-198"></a>
<div class="center"
>
<!--l. 137--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-111">
<div class="minipage"><pre class="verbatim" id="verbatim-113">
call&#x00A0;psb_cuda_exit(ctxt)
</pre>
<!--l. 141--><p class="nopar" > </div></div>
@@ -106,16 +106,16 @@ class="pplb7t-">in</span>.<br
class="newline" />Specified as: an integer variable.</dd></dl>
<!--l. 161--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-158000"></a>psb_cuda_DeviceSync &#8212; Synchronize CUDA device</h4>
id="x20-160000"></a>psb_cuda_DeviceSync &#8212; Synchronize CUDA device</h4>
<a
id="Q1-20-198"></a>
id="Q1-20-200"></a>
<div class="center"
>
<!--l. 169--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-112">
<div class="minipage"><pre class="verbatim" id="verbatim-114">
call&#x00A0;psb_cuda_DeviceSync()
</pre>
<!--l. 173--><p class="nopar" > </div></div>
@@ -123,26 +123,26 @@ call&#x00A0;psb_cuda_DeviceSync()
CUDA-side code, have completed.
<!--l. 182--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-159000"></a>psb_cuda_getDeviceCount </h4>
id="x20-161000"></a>psb_cuda_getDeviceCount </h4>
<a
id="Q1-20-200"></a>
id="Q1-20-202"></a>
<div class="center"
>
<!--l. 190--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-113">
<div class="minipage"><pre class="verbatim" id="verbatim-115">
ngpus&#x00A0;=&#x00A0;&#x00A0;psb_cuda_getDeviceCount()
</pre>
<!--l. 194--><p class="nopar" > </div></div>
<!--l. 199--><p class="noindent" >Get number of devices available on current computing node.
<!--l. 201--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-160000"></a>psb_cuda_getDevice </h4>
id="x20-162000"></a>psb_cuda_getDevice </h4>
<a
id="Q1-20-202"></a>
id="Q1-20-204"></a>
<div class="center"
>
<!--l. 209--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-114">
<div class="minipage"><pre class="verbatim" id="verbatim-116">
ngpus&#x00A0;=&#x00A0;&#x00A0;psb_cuda_getDevice()
</pre>
<!--l. 213--><p class="nopar" > </div></div>
@@ -152,39 +152,39 @@ ngpus&#x00A0;=&#x00A0;&#x00A0;psb_cuda_getDevice()
<!--l. 220--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-161000"></a>psb_cuda_setDevice </h4>
id="x20-163000"></a>psb_cuda_setDevice </h4>
<a
id="Q1-20-204"></a>
id="Q1-20-206"></a>
<div class="center"
>
<!--l. 228--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-115">
<div class="minipage"><pre class="verbatim" id="verbatim-117">
info&#x00A0;=&#x00A0;psb_cuda_setDevice(dev)
</pre>
<!--l. 232--><p class="nopar" > </div></div>
<!--l. 237--><p class="noindent" >Set device to be used by current process.
<!--l. 239--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-162000"></a>psb_cuda_DeviceHasUVA </h4>
id="x20-164000"></a>psb_cuda_DeviceHasUVA </h4>
<a
id="Q1-20-206"></a>
id="Q1-20-208"></a>
<div class="center"
>
<!--l. 247--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-116">
<div class="minipage"><pre class="verbatim" id="verbatim-118">
hasUva&#x00A0;=&#x00A0;psb_cuda_DeviceHasUVA()
</pre>
<!--l. 251--><p class="nopar" > </div></div>
<!--l. 256--><p class="noindent" >Returns true if device currently in use supports UVA (Unified Virtual Addressing).
<!--l. 259--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-163000"></a>psb_cuda_WarpSize </h4>
id="x20-165000"></a>psb_cuda_WarpSize </h4>
<a
id="Q1-20-208"></a>
id="Q1-20-210"></a>
<div class="center"
>
<!--l. 267--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-117">
<div class="minipage"><pre class="verbatim" id="verbatim-119">
nw&#x00A0;=&#x00A0;psb_cuda_WarpSize()
</pre>
<!--l. 271--><p class="nopar" > </div></div>
@@ -194,39 +194,39 @@ nw&#x00A0;=&#x00A0;psb_cuda_WarpSize()
<!--l. 279--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-164000"></a>psb_cuda_MultiProcessors </h4>
id="x20-166000"></a>psb_cuda_MultiProcessors </h4>
<a
id="Q1-20-210"></a>
id="Q1-20-212"></a>
<div class="center"
>
<!--l. 287--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-118">
<div class="minipage"><pre class="verbatim" id="verbatim-120">
nmp&#x00A0;=&#x00A0;psb_cuda_MultiProcessors()
</pre>
<!--l. 291--><p class="nopar" > </div></div>
<!--l. 296--><p class="noindent" >Returns the number of multiprocessors in the CUDA device.
<!--l. 298--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-165000"></a>psb_cuda_MaxThreadsPerMP </h4>
id="x20-167000"></a>psb_cuda_MaxThreadsPerMP </h4>
<a
id="Q1-20-212"></a>
id="Q1-20-214"></a>
<div class="center"
>
<!--l. 306--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-119">
<div class="minipage"><pre class="verbatim" id="verbatim-121">
nt&#x00A0;=&#x00A0;psb_cuda_MaxThreadsPerMP()
</pre>
<!--l. 310--><p class="nopar" > </div></div>
<!--l. 315--><p class="noindent" >Returns the maximum number of threads per multiprocessor.
<!--l. 318--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-166000"></a>psb_cuda_MaxRegistersPerBlock </h4>
id="x20-168000"></a>psb_cuda_MaxRegistersPerBlock </h4>
<a
id="Q1-20-214"></a>
id="Q1-20-216"></a>
<div class="center"
>
<!--l. 326--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-120">
<div class="minipage"><pre class="verbatim" id="verbatim-122">
nr&#x00A0;=&#x00A0;psb_cuda_MaxRegistersPerBlock()
</pre>
<!--l. 330--><p class="nopar" > </div></div>
@@ -236,39 +236,39 @@ nr&#x00A0;=&#x00A0;psb_cuda_MaxRegistersPerBlock()
<!--l. 338--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-167000"></a>psb_cuda_MemoryClockRate </h4>
id="x20-169000"></a>psb_cuda_MemoryClockRate </h4>
<a
id="Q1-20-216"></a>
id="Q1-20-218"></a>
<div class="center"
>
<!--l. 346--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-121">
<div class="minipage"><pre class="verbatim" id="verbatim-123">
cl&#x00A0;=&#x00A0;psb_cuda_MemoryClockRate()
</pre>
<!--l. 350--><p class="nopar" > </div></div>
<!--l. 355--><p class="noindent" >Returns the memory clock rate in KHz, as an integer.
<!--l. 357--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-168000"></a>psb_cuda_MemoryBusWidth </h4>
id="x20-170000"></a>psb_cuda_MemoryBusWidth </h4>
<a
id="Q1-20-218"></a>
id="Q1-20-220"></a>
<div class="center"
>
<!--l. 365--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-122">
<div class="minipage"><pre class="verbatim" id="verbatim-124">
nb&#x00A0;=&#x00A0;psb_cuda_MemoryBusWidth()
</pre>
<!--l. 369--><p class="nopar" > </div></div>
<!--l. 374--><p class="noindent" >Returns the memory bus width in bits.
<!--l. 376--><p class="noindent" >
<h4 class="likesubsectionHead"><a
id="x20-169000"></a>psb_cuda_MemoryPeakBandwidth </h4>
id="x20-171000"></a>psb_cuda_MemoryPeakBandwidth </h4>
<a
id="Q1-20-220"></a>
id="Q1-20-222"></a>
<div class="center"
>
<!--l. 384--><p class="noindent" >
<div class="minipage"><pre class="verbatim" id="verbatim-123">
<div class="minipage"><pre class="verbatim" id="verbatim-125">
bw&#x00A0;=&#x00A0;psb_cuda_MemoryPeakBandwidth()
</pre>
<!--l. 388--><p class="nopar" > </div></div>